MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641047826 A) filed by Srm Institute Of Science And Technology, Ramapuram Campus; and Easwari Engineering College, Chennai, Tamil Nadu, on April 15, for 'a deep learning-based real-time multi-object detection and motion-aware tracking system for intelligent video surveillance and dynamic environment monitoring detection.'
Inventor(s) include Dr. Adlene Ebenezer P; Dr P Sabitha; Aniket K S; Deepak R; Mikhil Sai N; and Mohammed Daanish.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "Object detection and tracking in video sequences play a crucial role in modern computer vision systems and support a wide range of applications such as intelligent surveillance, autonomous navigation, human-computer interaction, and advanced video analytics. Despite significant progress, achieving accurate and reliable tracking in dynamic environments remains a challenging task. Existing methods often face difficulties due to occlusion, variations in lighting, complex backgrounds, camera movement, and the need for real-time performance, which can negatively affect detection accuracy and tracking stability. This research aims to develop a robust and efficient framework for object detection and tracking that ensures accurate localization and consistent identity maintenance of objects across consecutive video frames while keeping computational complexity low. To achieve this, the proposed system combines deep learning-based object detection with motion-aware tracking techniques. A convolutional neural network (CNN) is utilized to detect objects in individual frames, and a tracking mechanism based on feature matching and motion prediction is employed to associate detected objects over time. The effectiveness of the proposed approach is evaluated using standard benchmark datasets and commonly used performance metrics, including precision, recall, and tracking accuracy. The scope of this study covers both real-time and offline video processing for single-object and multi-object tracking scenarios, making the framework suitable for applications such as traffic surveillance, security monitoring, and sports analysis. Experimental results indicate that the proposed method achieves higher detection accuracy, more stable tracking, and reduced identity switching compared to conventional approaches, demonstrating its effectiveness and practical applicability."
Disclaimer: Curated by HT Syndication.
